FACE RECOGNITION TECHNOLOGY JITHIN S J S 7

FACE RECOGNITION TECHNOLOGY JITHIN. S. J. S 7 CSE. 123 seminarsonly. com

FACE RECOGNITION All identification and authentication technologies operate using the 4 stage procedure Ø capture Ø extraction Ø comparison Ø match/nonmatch 123 seminarsonly. com

There are two methods of capture Ø Video imaging uses video cameras. System performance depends on precise position, angle of the head and surrounding lighting condition. Ø Thermal imaging has better accuracy as it uses facial temperature variation. Disadvantage is that its very costly. 123 seminarsonly. com

CAPTURING OF IMAGE BY STANDARD VIDEO CAMERAS 123 seminarsonly. com

COMPONENTS OF FACE RECONITION SYSTEM Consists of three modules Ø Enrollment module Ø Database Ø Identification 123 seminarsonly. com

Performance Parameters False acceptance rate, FAR= NFA/NIIA NFA= number of false acceptance NIIA=no of imposter identification attempt Ø False rejection rate, FRR = NFR/NEIA NFR=number of false rejection NEIA=number of enrollee identification attempt Ø Threshold/decision threshold Ø Equal error rate Ø 123 seminarsonly. com

IMPLEMENTATION OF FACE RECOGNITION TECHNOLOGY Ø Data acquisition Ø Input processing Ø Face image classification and decision making 123 seminarsonly. com

HOW FACE RECOGNITION SYSTEMS WORKS-an example Visionics, a company based in New Jersey Faceit There are distinguishable landmark on your face called peaks and valleys-nodal points Ø 80 nodal points on human face like -distance between eyes -width of nose -depth of eye socket -cheek bones -jawline -chin Ø faceit needs 14 to 20 nodal points Ø Ø Ø 123 seminarsonly. com

THE SOFTWARE The basic process used by Faceit system are Ø Detection Ø Alignment Ø Normalisation Ø Representation Ø Matching 123 seminarsonly. com

ADVANTAGES AND DISADVANTAGES Ø Convenient, social acceptability Ø Easy to use Ø Inexpensive technique of identification DISADVANTAGES Ø Problem with false rejection when people change their hair style, grow or shave a beard or wear glasses. Ø Identical twins 123 seminarsonly. com

WHY WE CHOOSE FACE RECOGNITION No physical interaction on behalf of the user Ø Does not require an expert to interpret the comparison Ø Can use existing hardware infrastructure Ø Can use existing image without having to reenroll everytime Ø Only biometric that allows you to perform passive identification in a one to many environment Ø 123 seminarsonly. com

APPLICATIONS Replacement of PIN, physical tokens Ø No need of human assistance for identification Ø Prison visitor systems Ø Border control Ø Voting system Ø Computer security Ø Banking using ATM Ø Physical access control of buildings, areas etc Ø 123 seminarsonly. com

CONCLUSION Ø Cost of equipments is going down-cost effective Ø Reliable Ø No technological or financial barriers 123 seminarsonly. com

Thank you 123 seminarsonly. com